3 * StatusNet - the distributed open-source microblogging tool
4 * Copyright (C) 2008, 2009, StatusNet, Inc.
6 * This program is free software: you can redistribute it and/or modify
7 * it under the terms of the GNU Affero General Public License as published by
8 * the Free Software Foundation, either version 3 of the License, or
9 * (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU Affero General Public License for more details.
16 * You should have received a copy of the GNU Affero General Public License
17 * along with this program. If not, see <http://www.gnu.org/licenses/>.
20 if (!defined('GNUSOCIAL')) { exit(1); }
23 * Table Definition for file
25 class File extends Managed_DataObject
27 public $__table = 'file'; // table name
28 public $id; // int(4) primary_key not_null
29 public $urlhash; // varchar(64) unique_key
31 public $filehash; // varchar(64) indexed
32 public $mimetype; // varchar(50)
33 public $size; // int(4)
34 public $title; // varchar(191) not 255 because utf8mb4 takes more space
35 public $date; // int(4)
36 public $protected; // int(4)
37 public $filename; // varchar(191) not 255 because utf8mb4 takes more space
38 public $width; // int(4)
39 public $height; // int(4)
40 public $modified; // timestamp() not_null default_CURRENT_TIMESTAMP
42 const URLHASH_ALG = 'sha256';
43 const FILEHASH_ALG = 'sha256';
45 public static function schemaDef()
49 'id' => array('type' => 'serial', 'not null' => true),
50 'urlhash' => array('type' => 'varchar', 'length' => 64, 'not null' => true, 'description' => 'sha256 of destination URL (url field)'),
51 'url' => array('type' => 'text', 'description' => 'destination URL after following possible redirections'),
52 'filehash' => array('type' => 'varchar', 'length' => 64, 'not null' => false, 'description' => 'sha256 of the file contents, only for locally stored files of course'),
53 'mimetype' => array('type' => 'varchar', 'length' => 50, 'description' => 'mime type of resource'),
54 'size' => array('type' => 'int', 'description' => 'size of resource when available'),
55 'title' => array('type' => 'varchar', 'length' => 191, 'description' => 'title of resource when available'),
56 'date' => array('type' => 'int', 'description' => 'date of resource according to http query'),
57 'protected' => array('type' => 'int', 'description' => 'true when URL is private (needs login)'),
58 'filename' => array('type' => 'varchar', 'length' => 191, 'description' => 'if a local file, name of the file'),
59 'width' => array('type' => 'int', 'description' => 'width in pixels, if it can be described as such and data is available'),
60 'height' => array('type' => 'int', 'description' => 'height in pixels, if it can be described as such and data is available'),
62 'modified' => array('type' => 'timestamp', 'not null' => true, 'description' => 'date this record was modified'),
64 'primary key' => array('id'),
65 'unique keys' => array(
66 'file_urlhash_key' => array('urlhash'),
69 'file_filehash_idx' => array('filehash'),
74 function isProtected($url) {
75 return 'http://www.facebook.com/login.php' === $url;
79 * Save a new file record.
81 * @param array $redir_data lookup data eg from File_redirection::where()
82 * @param string $given_url
85 public static function saveNew(array $redir_data, $given_url) {
87 // I don't know why we have to keep doing this but I'm adding this last check to avoid
90 $file = File::getKV('urlhash', self::hashurl($given_url));
92 if (!$file instanceof File) {
94 $file->urlhash = self::hashurl($given_url);
95 $file->url = $given_url;
96 if (!empty($redir_data['protected'])) $file->protected = $redir_data['protected'];
97 if (!empty($redir_data['title'])) $file->title = $redir_data['title'];
98 if (!empty($redir_data['type'])) $file->mimetype = $redir_data['type'];
99 if (!empty($redir_data['size'])) $file->size = intval($redir_data['size']);
100 if (isset($redir_data['time']) && $redir_data['time'] > 0) $file->date = intval($redir_data['time']);
101 $file_id = $file->insert();
104 Event::handle('EndFileSaveNew', array($file, $redir_data, $given_url));
105 assert ($file instanceof File);
110 * Go look at a URL and possibly save data about it if it's new:
111 * - follow redirect chains and store them in file_redirection
112 * - if a thumbnail is available, save it in file_thumbnail
113 * - save file record with basic info
114 * - optionally save a file_to_post record
115 * - return the File object with the full reference
117 * @fixme refactor this mess, it's gotten pretty scary.
118 * @param string $given_url the URL we're looking at
119 * @param Notice $notice (optional)
120 * @param bool $followRedirects defaults to true
122 * @return mixed File on success, -1 on some errors
124 * @throws ServerException on failure
126 public static function processNew($given_url, Notice $notice=null, $followRedirects=true) {
127 if (empty($given_url)) {
128 throw new ServerException('No given URL to process');
131 $given_url = File_redirection::_canonUrl($given_url);
132 if (empty($given_url)) {
133 throw new ServerException('No canonical URL from given URL to process');
139 $file = File::getByUrl($given_url);
140 } catch (NoResultException $e) {
141 // First check if we have a lookup trace for this URL already
143 $file_redir = File_redirection::getByUrl($given_url);
144 $file = File::getKV('id', $file_redir->file_id);
145 if (!$file instanceof File) {
146 // File did not exist, let's clean up the File_redirection entry
147 $file_redir->delete();
149 } catch (NoResultException $e) {
150 // We just wanted to doublecheck whether a File_thumbnail we might've had
151 // actually referenced an existing File object.
155 // If we still don't have a File object, let's create one now!
156 if (!$file instanceof File) {
157 // @fixme for new URLs this also looks up non-redirect data
158 // such as target content type, size, etc, which we need
159 // for File::saveNew(); so we call it even if not following
161 $redir_data = File_redirection::where($given_url);
162 if (is_array($redir_data)) {
163 $redir_url = $redir_data['url'];
164 } elseif (is_string($redir_data)) {
165 $redir_url = $redir_data;
166 $redir_data = array();
168 // TRANS: Server exception thrown when a URL cannot be processed.
169 throw new ServerException(sprintf(_("Cannot process URL '%s'"), $given_url));
172 if ($redir_url === $given_url || !$followRedirects) {
173 // Save the File object based on our lookup trace
174 $file = File::saveNew($redir_data, $given_url);
176 // This seems kind of messed up... for now skipping this part
177 // if we're already under a redirect, so we don't go into
178 // horrible infinite loops if we've been given an unstable
179 // redirect (where the final destination of the first request
180 // doesn't match what we get when we ask for it again).
182 // Seen in the wild with clojure.org, which redirects through
183 // wikispaces for auth and appends session data in the URL params.
184 $file = self::processNew($redir_url, $notice, /*followRedirects*/false);
185 File_redirection::saveNew($redir_data, $file->id, $given_url);
188 if (!$file instanceof File) {
189 // This should only happen if File::saveNew somehow did not return a File object,
190 // though we have an assert for that in case the event there might've gone wrong.
191 // If anything else goes wrong, there should've been an exception thrown.
192 throw new ServerException('URL processing failed without new File object');
196 if ($notice instanceof Notice) {
197 File_to_post::processNew($file, $notice);
202 public static function respectsQuota(Profile $scoped, $fileSize) {
203 if ($fileSize > common_config('attachments', 'file_quota')) {
204 // TRANS: Message used to be inserted as %2$s in the text "No file may
205 // TRANS: be larger than %1$d byte and the file you sent was %2$s.".
206 // TRANS: %1$d is the number of bytes of an uploaded file.
207 $fileSizeText = sprintf(_m('%1$d byte','%1$d bytes',$fileSize),$fileSize);
209 $fileQuota = common_config('attachments', 'file_quota');
210 // TRANS: Message given if an upload is larger than the configured maximum.
211 // TRANS: %1$d (used for plural) is the byte limit for uploads,
212 // TRANS: %2$s is the proper form of "n bytes". This is the only ways to have
213 // TRANS: gettext support multiple plurals in the same message, unfortunately...
214 throw new ClientException(
215 sprintf(_m('No file may be larger than %1$d byte and the file you sent was %2$s. Try to upload a smaller version.',
216 'No file may be larger than %1$d bytes and the file you sent was %2$s. Try to upload a smaller version.',
218 $fileQuota, $fileSizeText));
223 $query = "select sum(size) as total from file join file_to_post on file_to_post.file_id = file.id join notice on file_to_post.post_id = notice.id where profile_id = {$scoped->id} and file.url like '%/notice/%/file'";
224 $file->query($query);
226 $total = $file->total + $fileSize;
227 if ($total > common_config('attachments', 'user_quota')) {
228 // TRANS: Message given if an upload would exceed user quota.
229 // TRANS: %d (number) is the user quota in bytes and is used for plural.
230 throw new ClientException(
231 sprintf(_m('A file this large would exceed your user quota of %d byte.',
232 'A file this large would exceed your user quota of %d bytes.',
233 common_config('attachments', 'user_quota')),
234 common_config('attachments', 'user_quota')));
236 $query .= ' AND EXTRACT(month FROM file.modified) = EXTRACT(month FROM now()) and EXTRACT(year FROM file.modified) = EXTRACT(year FROM now())';
237 $file->query($query);
239 $total = $file->total + $fileSize;
240 if ($total > common_config('attachments', 'monthly_quota')) {
241 // TRANS: Message given id an upload would exceed a user's monthly quota.
242 // TRANS: $d (number) is the monthly user quota in bytes and is used for plural.
243 throw new ClientException(
244 sprintf(_m('A file this large would exceed your monthly quota of %d byte.',
245 'A file this large would exceed your monthly quota of %d bytes.',
246 common_config('attachments', 'monthly_quota')),
247 common_config('attachments', 'monthly_quota')));
252 public function getFilename()
254 if (!self::validFilename($this->filename)) {
255 // TRANS: Client exception thrown if a file upload does not have a valid name.
256 throw new ClientException(_("Invalid filename."));
258 return $this->filename;
261 // where should the file go?
263 static function filename(Profile $profile, $origname, $mimetype)
265 $ext = self::guessMimeExtension($mimetype);
267 // Normalize and make the original filename more URL friendly.
268 $origname = basename($origname, ".$ext");
269 if (class_exists('Normalizer')) {
270 // http://php.net/manual/en/class.normalizer.php
271 // http://www.unicode.org/reports/tr15/
272 $origname = Normalizer::normalize($origname, Normalizer::FORM_KC);
274 $origname = preg_replace('/[^A-Za-z0-9\.\_]/', '_', $origname);
276 $nickname = $profile->getNickname();
277 $datestamp = strftime('%Y%m%d', time());
279 // generate new random strings until we don't run into a filename collision.
280 $random = strtolower(common_confirmation_code(16));
281 $filename = "$nickname-$datestamp-$origname-$random.$ext";
282 } while (file_exists(self::path($filename)));
286 static function guessMimeExtension($mimetype)
289 $ext = common_supported_mime_to_ext($mimetype);
290 } catch (Exception $e) {
291 // We don't support this mimetype, but let's guess the extension
292 $ext = substr(strrchr($mimetype, '/'), 1);
294 return strtolower($ext);
298 * Validation for as-saved base filenames
300 static function validFilename($filename)
302 return preg_match('/^[A-Za-z0-9._-]+$/', $filename);
306 * @throws ClientException on invalid filename
308 static function path($filename)
310 if (!self::validFilename($filename)) {
311 // TRANS: Client exception thrown if a file upload does not have a valid name.
312 throw new ClientException(_("Invalid filename."));
314 $dir = common_config('attachments', 'dir');
316 if ($dir[strlen($dir)-1] != '/') {
320 return $dir . $filename;
323 static function url($filename)
325 if (!self::validFilename($filename)) {
326 // TRANS: Client exception thrown if a file upload does not have a valid name.
327 throw new ClientException(_("Invalid filename."));
330 if (common_config('site','private')) {
332 return common_local_url('getfile',
333 array('filename' => $filename));
337 if (GNUsocial::useHTTPS()) {
339 $sslserver = common_config('attachments', 'sslserver');
341 if (empty($sslserver)) {
342 // XXX: this assumes that background dir == site dir + /file/
343 // not true if there's another server
344 if (is_string(common_config('site', 'sslserver')) &&
345 mb_strlen(common_config('site', 'sslserver')) > 0) {
346 $server = common_config('site', 'sslserver');
347 } else if (common_config('site', 'server')) {
348 $server = common_config('site', 'server');
350 $path = common_config('site', 'path') . '/file/';
352 $server = $sslserver;
353 $path = common_config('attachments', 'sslpath');
355 $path = common_config('attachments', 'path');
361 $path = common_config('attachments', 'path');
362 $server = common_config('attachments', 'server');
364 if (empty($server)) {
365 $server = common_config('site', 'server');
368 $ssl = common_config('attachments', 'ssl');
370 $protocol = ($ssl) ? 'https' : 'http';
373 if ($path[strlen($path)-1] != '/') {
377 if ($path[0] != '/') {
381 return $protocol.'://'.$server.$path.$filename;
384 function getEnclosure(){
385 $enclosure = (object) array();
386 foreach (array('title', 'url', 'date', 'modified', 'size', 'mimetype') as $key) {
387 $enclosure->$key = $this->$key;
390 $needMoreMetadataMimetypes = array(null, 'application/xhtml+xml');
392 if (!isset($this->filename) && in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) {
393 // This fetches enclosure metadata for non-local links with unset/HTML mimetypes,
394 // which may be enriched through oEmbed or similar (implemented as plugins)
395 Event::handle('FileEnclosureMetadata', array($this, &$enclosure));
397 if (empty($enclosure->mimetype) || in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) {
398 // This means we either don't know what it is, so it can't
399 // be shown as an enclosure, or it is an HTML link which
400 // does not link to a resource with further metadata.
401 throw new ServerException('Unknown enclosure mimetype, not enough metadata');
407 * Get the attachment's thumbnail record, if any.
408 * Make sure you supply proper 'int' typed variables (or null).
410 * @param $width int Max width of thumbnail in pixels. (if null, use common_config values)
411 * @param $height int Max height of thumbnail in pixels. (if null, square-crop to $width)
412 * @param $crop bool Crop to the max-values' aspect ratio
414 * @return File_thumbnail
416 * @throws UseFileAsThumbnailException if the file is considered an image itself and should be itself as thumbnail
417 * @throws UnsupportedMediaException if, despite trying, we can't understand how to make a thumbnail for this format
418 * @throws ServerException on various other errors
420 public function getThumbnail($width=null, $height=null, $crop=false, $force_still=true)
422 // Get some more information about this file through our ImageFile class
423 $image = ImageFile::fromFileObject($this);
424 if ($image->animated && !common_config('thumbnail', 'animated')) {
425 // null means "always use file as thumbnail"
426 // false means you get choice between frozen frame or original when calling getThumbnail
427 if (is_null(common_config('thumbnail', 'animated')) || !$force_still) {
428 throw new UseFileAsThumbnailException($this->id);
432 return $image->getFileThumbnail($width, $height, $crop);
435 public function getPath()
437 $filepath = self::path($this->filename);
438 if (!file_exists($filepath)) {
439 throw new FileNotFoundException($filepath);
444 public function getUrl()
446 if (!empty($this->filename)) {
447 // A locally stored file, so let's generate a URL for our instance.
448 $url = self::url($this->filename);
449 if (self::hashurl($url) !== $this->urlhash) {
450 // For indexing purposes, in case we do a lookup on the 'url' field.
451 // also we're fixing possible changes from http to https, or paths
452 $this->updateUrl($url);
457 // No local filename available, return the URL we have stored
461 static public function getByUrl($url)
464 $file->urlhash = self::hashurl($url);
465 if (!$file->find(true)) {
466 throw new NoResultException($file);
472 * @param string $hashstr String of (preferrably lower case) hexadecimal characters, same as result of 'hash_file(...)'
474 static public function getByHash($hashstr, $alg=File::FILEHASH_ALG)
477 $file->filehash = strtolower($hashstr);
478 if (!$file->find(true)) {
479 throw new NoResultException($file);
484 public function updateUrl($url)
486 $file = File::getKV('urlhash', self::hashurl($url));
487 if ($file instanceof File) {
488 throw new ServerException('URL already exists in DB');
490 $sql = 'UPDATE %1$s SET urlhash=%2$s, url=%3$s WHERE urlhash=%4$s;';
491 $result = $this->query(sprintf($sql, $this->tableName(),
492 $this->_quote((string)self::hashurl($url)),
493 $this->_quote((string)$url),
494 $this->_quote((string)$this->urlhash)));
495 if ($result === false) {
496 common_log_db_error($this, 'UPDATE', __FILE__);
497 throw new ServerException("Could not UPDATE {$this->tableName()}.url");
504 * Blow the cache of notices that link to this URL
506 * @param boolean $last Whether to blow the "last" cache too
511 function blowCache($last=false)
513 self::blow('file:notice-ids:%s', $this->id);
515 self::blow('file:notice-ids:%s;last', $this->id);
517 self::blow('file:notice-count:%d', $this->id);
521 * Stream of notices linking to this URL
523 * @param integer $offset Offset to show; default is 0
524 * @param integer $limit Limit of notices to show
525 * @param integer $since_id Since this notice
526 * @param integer $max_id Before this notice
528 * @return array ids of notices that link to this file
531 function stream($offset=0, $limit=NOTICES_PER_PAGE, $since_id=0, $max_id=0)
533 $stream = new FileNoticeStream($this);
534 return $stream->getNotices($offset, $limit, $since_id, $max_id);
537 function noticeCount()
539 $cacheKey = sprintf('file:notice-count:%d', $this->id);
541 $count = self::cacheGet($cacheKey);
543 if ($count === false) {
545 $f2p = new File_to_post();
547 $f2p->file_id = $this->id;
549 $count = $f2p->count();
551 self::cacheSet($cacheKey, $count);
557 public function isLocal()
559 return !empty($this->filename);
562 public function delete($useWhere=false)
564 // Delete the file, if it exists locally
565 if (!empty($this->filename) && file_exists(self::path($this->filename))) {
566 $deleted = @unlink(self::path($this->filename));
568 common_log(LOG_ERR, sprintf('Could not unlink existing file: "%s"', self::path($this->filename)));
572 // Clear out related things in the database and filesystem, such as thumbnails
573 if (Event::handle('FileDeleteRelated', array($this))) {
574 $thumbs = new File_thumbnail();
575 $thumbs->file_id = $this->id;
576 if ($thumbs->find()) {
577 while ($thumbs->fetch()) {
582 $f2p = new File_to_post();
583 $f2p->file_id = $this->id;
585 while ($f2p->fetch()) {
591 // And finally remove the entry from the database
592 return parent::delete($useWhere);
595 public function getTitle()
597 $title = $this->title ?: $this->filename;
599 return $title ?: null;
602 static public function hashurl($url)
605 throw new Exception('No URL provided to hash algorithm.');
607 return hash(self::URLHASH_ALG, $url);
610 static public function beforeSchemaUpdate()
612 $table = strtolower(get_called_class());
613 $schema = Schema::get();
614 $schemadef = $schema->getTableDef($table);
616 // 2015-02-19 We have to upgrade our table definitions to have the urlhash field populated
617 if (isset($schemadef['fields']['urlhash']) && isset($schemadef['unique keys']['file_urlhash_key'])) {
618 // We already have the urlhash field, so no need to migrate it.
621 echo "\nFound old $table table, upgrading it to contain 'urlhash' field...";
624 $file->query(sprintf('SELECT id, LEFT(url, 191) AS shortenedurl, COUNT(*) AS c FROM %1$s WHERE LENGTH(url)>191 GROUP BY shortenedurl HAVING c > 1', $schema->quoteIdentifier($table)));
625 print "\nFound {$file->N} URLs with too long entries in file table\n";
626 while ($file->fetch()) {
627 // We've got a URL that is too long for our future file table
628 // so we'll cut it. We could save the original URL, but there is
629 // no guarantee it is complete anyway since the previous max was 255 chars.
630 $dupfile = new File();
631 // First we find file entries that would be duplicates of this when shortened
632 // ... and we'll just throw the dupes out the window for now! It's already so borken.
633 $dupfile->query(sprintf('SELECT * FROM file WHERE LEFT(url, 191) = "%1$s"', $file->shortenedurl));
634 // Leave one of the URLs in the database by using ->find(true) (fetches first entry)
635 if ($dupfile->find(true)) {
636 print "\nShortening url entry for $table id: {$file->id} [";
637 $orig = clone($dupfile);
638 $dupfile->url = $file->shortenedurl; // make sure it's only 191 chars from now on
639 $dupfile->update($orig);
640 print "\nDeleting duplicate entries of too long URL on $table id: {$file->id} [";
641 // only start deleting with this fetch.
642 while($dupfile->fetch()) {
648 print "\nWarning! URL suddenly disappeared from database: {$file->url}\n";
651 echo "...and now all the non-duplicates which are longer than 191 characters...\n";
652 $file->query('UPDATE file SET url=LEFT(url, 191) WHERE LENGTH(url)>191');
654 echo "\n...now running hacky pre-schemaupdate change for $table:";
655 // We have to create a urlhash that is _not_ the primary key,
656 // transfer data and THEN run checkSchema
657 $schemadef['fields']['urlhash'] = array (
660 'not null' => false, // this is because when adding column, all entries will _be_ NULL!
661 'description' => 'sha256 of destination URL (url field)',
663 $schemadef['fields']['url'] = array (
665 'description' => 'destination URL after following possible redirections',
667 unset($schemadef['unique keys']);
668 $schema->ensureTable($table, $schemadef);
671 $classname = ucfirst($table);
672 $tablefix = new $classname;
673 // urlhash is hash('sha256', $url) in the File table
674 echo "Updating urlhash fields in $table table...";
675 // Maybe very MySQL specific :(
676 $tablefix->query(sprintf('UPDATE %1$s SET %2$s=%3$s;',
677 $schema->quoteIdentifier($table),
679 // The line below is "result of sha256 on column `url`"
682 echo "Resuming core schema upgrade...";